raloxifene: risks and benefits

295

Raloxifene: Risks and Benefits

ELIZABETH BARRETT-CONNOR

Division of Epidemiology, Department of Family and Preventive Medicine,University of California, San Diego, La Jolla, California 92093, USA

ABSTRACT: Raloxifene, a selective estrogen receptor modulator (SERM), wasdesigned to have the expected benefits of long-term estrogen replacement ther-apy without the risks. This paper reviews the clinical trial evidence for ralox-ifene benefits and risks, and how they compare with those of hormonereplacement therapy (HRT) and relate to the choices of postmenopausalwomen.

KEYWORDS: selective estrogen receptor modulators (SERMs); raloxifene;risks; benefits

INTRODUCTION

Estrogen receptors bind with multiple ligands, eliciting tissue-specific responses.Selective estrogen receptor modulators (SERMs) are designed to have both estrogenagonist and antagonist effects. Raloxifene is the first SERM approved for osteoporo-sis prevention and treatment. This paper summarizes the evidence for raloxifene’sbenefits and risks based on results from randomized placebo-controlled clinical tri-als. Unless otherwise indicated, only data related to the now-recommended dose,60 mg/day, are reviewed.

BENEFITS

Osteoporosis Prevention and Treatment

Osteoporosis is a major cause of morbidity and mortality in elderly women. Thelargest completed clinical trial of raloxifene and fractures is the Multiple Outcomes ofRaloxifene Evaluation (MORE), in which 5,153 postmenopausal women were ran-domized to either placebo or raloxifene at the 60 mg daily dose.1 At baseline, all par-ticipants (mean age 67 years) had osteoporosis as defined by bone mineral density orprior vertebral fracture criteria. At 36 months of treatment, raloxifene had no signifi-cant effect on the risk of nonvertebral fractures, but the risk of radiographic vertebralfractures was reduced significantly in women with and without prevalent fractures atbaseline. As shown in FIGURE 1, the 3-year risk of a vertebral fracture in women with-out prior spine fractures was reduced by 50% (4.5% on placebo and 2.3% on ralox-ifene); for women with prior vertebral fractures, the risk of new vertebral fractures was

Address for correspondence: Elizabeth Barrett-Connor, M.D., Professor of Family and Pre-ventive Medicine, Chief, Division of Epidemiology, 9500 Gilman Drive #0607, University ofCalifornia, San Diego, La Jolla, CA 92093.

[email protected]

296 ANNALS NEW YORK ACADEMY OF SCIENCES

reduced by 30% (21.2% on placebo and 14.7% on raloxifene). Raloxifene reduced therisk of symptomatic vertebral fractures by 61% within the first year.2

Breast Cancer

Breast cancer is the most common cause of cancer death in women worldwide. Inthe MORE trial,3 invasive breast cancers were reduced by 76% in women assigned toraloxifene (3.6 cancers per 1,000 woman-years in the placebo group vs. 0.9 cancers per1,000 woman-years in the pooled raloxifene 60 mg/day and 120 mg/day group).Estrogen receptor (ER) positive cancers were reduced by 90%, as shown in FIGURE 2.The overall reduction in breast cancer risk was maintained for at least 48 months.4 Al-though MORE women were selected for osteoporosis, and might therefore be expectedto be at a reduced risk of breast cancer, breast cancer incidence in the MORE placebogroup was similar to that expected in the general population of the same age.

Because the observed breast cancer risk reduction in MORE was based on fewcases and was not a planned primary outcome, the trial has been extended with a newname, Continuing Outcomes Relevant to Evista (CORE), with invasive breast cancer

FIGURE 2. Estrogen receptor (ER) positive breast cancers were reduced by 90% inwomen assigned to raloxifene. (Adapted from Cummings et al.3)

FIGURE 1. The risk of vertebral fracture in women without prior spine fractures was re-duced by 50% and by 30% for women with prior spine fractures. (Adapted from Ettinger et al.1)

297BARRETT-CONNOR: RISKS AND BENEFITS OF RALOXIFENE

as a designated outcome. CORE will enroll approximately 4,000 MORE participantsfor an additional four years, extending the total trial duration to eight years. Invasivebreast cancer is also a co-primary outcome in the Raloxifene Use and The Heart(RUTH) trial (see below) and in the Study of Tamoxifen and Raloxifene (STAR).5

The latter will compare raloxifene with tamoxifen in 22,000 postmenopausal womenat high risk for breast cancer.

Endometrial Bleeding, Hyperplasia, and Cancer

In the MORE trial, bleeding rates were the same in women assigned to raloxifene(3.1%) or placebo (3.1%).3 Fluid in the endometrial cavity (observed on transvaginalultrasound) was present in more women assigned to raloxifene (2.4%) than placebo;the clinical significance of this finding is unknown.

In clinical trials of up to five years’ duration, there were very low rates of en-dometrial hyperplasia and cancer in women treated with raloxifene or placebo, withno significant differences between treatment groups.3,6–8 The STAR breast cancerprevention trial is also designed to determine whether raloxifene reduces the risk ofendometrial cancer.

Cardiovascular Disease

Heart disease is the most common cause of death in women in most of the westernworld. For this reason, the largest potential benefit for a SERM is the prevention ofcoronary heart disease. No clinical trial data show an altered risk of heart disease inwomen taking raloxifene. The large clinical trial Raloxifene Use for The Heart(RUTH) was designed to determine whether raloxifene will reduce the risk of coro-nary heart disease in women at high risk of heart disease. Among the 10,101 womenenrolled, about half have known coronary heart disease and about half have heart dis-ease risk factors, most commonly diabetes. In RUTH, the primary cardiovascularend point is a combined outcome of myocardial infarction, coronary heart disease–related death, or hospitalization for acute coronary syndrome. The rationale for thisstudy was the observation that raloxifene improves several heart disease risk factors,lowering LDL cholesterol and apolipoprotein B level without raising triglycerides,and lowering homocysteine and fibrinogen levels.9,10

RISKS

Venous Thromboembolic Events

To date, the only serious adverse event clearly associated with raloxifene is a tri-pling of the risk of venous thromboembolic events (VTE).3 By 40 months of followup, deep vein thrombosis had occurred in 7/1,000 MORE women assigned to ralox-ifene versus 2/1,000 assigned to placebo, and pulmonary embolus in 3/1,000 as-signed to raloxifene versus 1/1,000 assigned to placebo.

Hot Flashes

Raloxifene increases hot flashes. In pooled analyses of trial data including 1,165younger postmenopausal women (average age 55 years), those assigned to ralox-


ifene reported more hot flashes (24.6%) than did women assigned to placebo(18.3%).11 In similar comparisons in the older MORE women (average age 67years), hot flashes were much less common but still reported by more women takingraloxifene (9.7%) than by women taking placebo (6.4%).1 In both age groups, hotflashes tended to be mild and led to discontinuation of raloxifene by less than 2%,12

but trials excluded women with severe vasomotor symptoms.

Miscellaneous Adverse Events

Leg cramps were more common in MORE women assigned to raloxifene thanplacebo (7.0 vs. 3.7%).3 Compared to women assigned to placebo, MORE womenassigned to raloxifene also reported a significantly higher incidence of an influenza-like syndrome (2.1%); peripheral edema (0.8%); and worsening of diabetes (0.7%).3

Central Nervous System

In a 12-month randomized clinical trial of 143 women (mean age 68 years), theincidence of depressive symptoms, emotional lability, insomnia, anxiety, dizziness,malaise, and memory loss did not differ between women assigned to raloxifene orplacebo.13 In the MORE trial, overall performance on cognitive function tests did notdiffer between postmenopausal women assigned to raloxifene or placebo for threeyears. In an age-stratified analysis, however, women aged 70 or older performed bet-ter on two of six tests after taking raloxifene than did women taking placebo.14

RALOXIFENE VERSUS ESTROGEN

Osteoporosis. No head-to-head comparisons of estrogen versus raloxifene havebeen published. In the Postmenopausal Estrogen/Progestin Intervention (PEPI) trial,there was an average 4–5% increase in vertebral bone density after two years ofHRT,15 more than the 2–3% two-year increase with raloxifene.16 No large clinicaltrials of estrogen designed to evaluate fracture risk in osteoporotic women have beenreported. The Heart and Estrogen/Progestin Replacement Study (HERS), the largestclinical trial with nearly complete follow-up, studied women selected for heart dis-ease, not low bone density.17 Nevertheless, there were more than 250 new clinicalfractures during the trial and no difference in clinical fracture rates between HRT-treated and untreated women. Unfortunately, vertebral X-rays were not obtained, soa reduction in subclinical spine fractures could have been missed.

Breast Cancer. Unlike raloxifene, there are no clinical trial data showing an in-creased or decreased risk of breast cancer after hormone replacement therapy (HRT).The 33% increase observed in HERS women on hormones was not statistically dif-ferent from the placebo rate.17 Data from the PEPI trial have shown that HRT in-creases breast pain and breast density on mammography, with a higher incidenceafter combined therapy than with estrogen alone (about 45 vs. 15%).18 Radiographicbreast density is a risk factor for breast cancer.19 Unlike estrogen, raloxifene doesnot increase breast pain11 or breast density.20

Endometrial Pathology. Unlike raloxifene, HRT in standard doses commonlycauses bleeding during the first year of treatment; lower doses appear to reduce the


frequency and duration of bleeding. Unopposed estrogen caused endometrial hyper-plasia at a rate of 10% per year in PEPI women.21

Heart Disease. The effects of raloxifene and estrogen on heart disease risk factorshave been studied in clinical trials with head-to-head comparisons. As shown in FIG-URE 3, raloxifene and estrogen (HRT) similarly improved total and LDL cholesterol

FIGURE 3. Effects of raloxifene and estrogen (HRT) on heart disease risk factors: ral-oxifene was better than estrogen in lowering fibrinogen and apolipoprotein B (apo B) levels,and in not raising triglycerides (Trigs) or C reactive protein (CRP); only estrogen reducedLp (a) and PAI-1 and raised HDL cholesterol (HDL-C) and apolipoprotein A-1 (apo A1).(Adapted from Walsh et al.9,10)


and homocysteine levels. Raloxifene was better than estrogen in lowering fibrinogenand apolipoprotein B levels, and in not raising triglycerides or C reactive protein, butonly estrogen was effective in reducing Lp(a) and PAI-1, and in raising HDL choles-terol and apolipoprotein A-1.9–10,22 Changes in heart disease risk factors do not nec-essarily translate to favorable changes in the risk of heart disease, as illustrated bythe failure of HRT to reduce CHD or stroke risk in primary and secondary preventiontrials.17,23–26

Cognitive Function. Despite several small clinical trials showing that estrogenimproved cognition in young women soon after bilateral oophorectomy, no largeclinical trials have shown improved mood or cognition in older women withoutmenopausal symptoms. In HERS, cognitive function test performance was slightlypoorer in women on HRT compared to placebo. In another clinical trial, when wom-en with early dementia were assigned to estrogen, they scored more poorly on theClinical Dementia Rating Scale than did women assigned to placebo.27

Venous Thromboembolic Disease. Based on a comparison between women inHERS and women in MORE, the risk of VTE with raloxifene is similar to that withHRT. These risks are also similar to those found in observational studies of womennot selected for osteoporosis or CHD risk (FIG. 4).28,29

Menopause Symptoms. Estrogen is the treatment of choice for vasomotor symp-toms and urogenital dryness, in contrast to raloxifene which may worsen hot flashesand has no effect on vaginal dryness. In a randomized trial of 187 women with vag-inal atrophy, raloxifene did not reduce the beneficial effects of vaginal estrogencream or moisturizer.30,31

DISCUSSION

Weighing risks and benefits is necessary if we are to help patients make intelli-gent choices. Women can understand that their risk of breast cancer without hor-mone therapy varies based on age, body size, reproductive history, family history,and other factors. Yet relative and absolute risk can be confusing. Consider a disease

FIGURE 4. Comparing women in HERS and in MORE, the risk of VTE (venous throm-boembolic disease) with raloxifene is similar to that with HRT (hormone replacement therapy).


that affects 2/1,000 people without treatment and 4/1,000 with treatment—the rela-tive risk is 2 but the absolute excess risk is only 2/1,000 treated people. Absolute riskis preferable to relative risk, because only the former estimates the probability of thecondition without the intervention. On average, using hormone therapy for 15 yearsis associated with a relative risk of breast cancer of 1.5 (a 50% increase in risk), butadds only an additional 5–20 cases per 1,000 women.32

The absolute risk can be translated into the number needed to treat (NNT). Forexample, based on the data shown in FIGURE 1, 46 women without prevalent frac-tures would need to be treated for 36 months to prevent one vertebral fracture, com-pared to 16 women with prevalent vertebral fractures. FIGURE 1 also illustrates oneof the problems with NNT: despite the smaller number of women with prevalentfractures who need to be treated to prevent one new fracture, the number of fracturesoccurring despite treatment is larger in this group than in women with no fracturesat baseline. Thus, more women will suffer new fractures despite treatment, if onewaits for a vertebral fracture as an indication to treat.

Similar calculations can be made for breast cancer. Based on MORE data, 126women unselected for breast cancer risk would need to be treated to prevent onebreast cancer. No data are yet available for women at particularly high risk for breastcancer. Calculations cannot be made for cardiovascular disease: the MORE data sug-gest no benefit (or risk) for women with osteoporosis (presumably at low risk ofheart disease), but estimates for high-risk women will have to wait for the RUTHresults.

Finally, with regard to serious risk, the best estimate is that one excess VTE willoccur for every 155 women treated with raloxifene for three years.

The idea of absolute risk with and without treatment is relatively easy to under-stand. But none of the numbers, relative risk, absolute risk, or NNT, deal with indi-vidual women and their personal concerns. All medications have side effects, whichmay be more or less unacceptable to different women. Women who are most con-cerned about breast cancer will rarely accept a medication thought to increase breastcancer risk—even if the drug is shown to reduce the risk of a more common condi-tion such as heart disease. Women’s decisions are also influenced by their age andtheir personal experience. Women less than 60 are more likely to have friends withbreast cancer than heart disease or osteoporosis. Women with symptoms are moreapt to initiate and continue HRT, trading demonstrated current well-being for futurepossible risk. Social class is another determinant of medication initiation and con-tinuation—even when the medicine is provided at no cost. It may be that those witha weekly paycheck will be less likely to begin or continue medication, with possiblebenefits in the far future, than women with a more stable income and (presumably)more predictable future.

Finally, risk benefit estimates from clinical trials may be misleading. Althoughclinical trials are needed to document the efficacy of a new drug, they cannot guar-antee these findings will apply in clinical practice. Poor adherence by patients andphysicians can undermine the benefits demonstrated in clinical trials. And some-times post-marketing surveillance is necessary to see a rare but serious side effect.

At present, the following conclusions about raloxifene risks and benefits can bemade. Raloxifene has effects on several heart disease risk factors—in some instancesmore favorable than estrogen and in others less favorable. Whether raloxifene canreduce heart disease risk is being studied now. Because heart disease is such a com-


mon cause of morbidity and mortality in older women, heart disease preventionwould greatly enhance the risk-benefit ratio for raloxifene. However, given the doc-umented fracture prevention, and if the breast cancer risk reduction is confirmed, theoverall risk-benefit ratio for raloxifene should be favorable even if raloxifene doesnot reduce the risk of heart disease. An unfavorable risk-benefit ratio would emergeonly if raloxifene causes an early excess risk of cardiovascular disease, as has beenobserved with HRT.

REFERENCES

1. ETTINGER, B., D.M. BLACK, B.H. MITLAK, et al. 1999. Reduction of vertebral fracturerisk in postmenopausal women with osteoporosis treated with raloxifene: resultsfrom a 3-year randomized clinical trial. Multiple Outcomes of Raloxifene Evaluation(MORE) Investigators. JAMA 282: 637–645.

2. EASTELL, R., J. ADACHI, K. HARPER, et al. 2000. The effects of raloxifene on incidentvertebral fractures in postmenopausal women with osteoporosis: 4-year results fromthe MORE trial. Abstr. 492. The American Society of Bone and Mineral Research22nd Annual Meeting. Toronto, Canada.

3. CUMMINGS, S.R., S. ECKERT, K.A. KRUEGER, et al. 1999. The effect of raloxifene onrisk of breast cancer in postmenopausal women: results from the MORE randomizedtrial. Multiple Outcomes of Raloxifene Evaluation. JAMA 281: 2189–2197.

4. CAULEY, J., N. NORTON, M.E. LIPPMAN, et al. 2001. Continued breast cancer riskreduction in postmenopausal women treated with raloxifene: 4-year results from theMORE trial. Breast Cancer Res. Treat. 65: 125–134.

5. JORDAN, V.C. & M. MORROW. 1999. Tamoxifen, raloxifene, and the prevention ofbreast cancer. Endocr. Rev. 20: 253–278.

6. COHEN, F.J., S. WATTS, A. SHAH, et al. 2000. Uterine effects of 3-year raloxifene ther-apy in postmenopausal women younger than age 60. Obstet. Gynecol. 95: 104–110.

7. GOLDSTEIN, S.R., W.H. SCHEELE, S.K. RAJAGOPALAN, et al. 2000. A 12-month compar-ative study of raloxifene, estrogen, and placebo on the postmenopausalendometrium. Obstet. Gynecol. 95: 95–103.

8. JOLLY, E.E., J.T. O’GORMAN, R.J. AKERS, et al. 2000. Effect of raloxifene therapy for 5years on endometrial hyperplasia and endometrial cancer in postmenopausal womenunder the age of 60. Abstr. 2318. The Endocrine Society’s 82nd Annual Meeting,Toronto, Canada.

9. WALSH, B.W., L.H. KULLER, R.A. WILD, et al. 1998. Effects of raloxifene on serum lip-ids and coagulation factors in healthy postmenopausal women. JAMA 279: 1445–1451.

10. WALSH, B.W., S. PAUL, R.A. WILD, et al. 2000. The effects of hormone replacementtherapy and raloxifene on C-reactive protein and homocysteine in healthy postmeno-pausal women: a randomized, controlled trial. J. Clin. Endocrinol. Metab. 85: 214–218.

11. DAVIES, G.C., W.J. HUSTER, Y. LU, et al. 1999. Adverse events reported by postmeno-pausal women in controlled trials with raloxifene. Obstet. Gynecol. 93: 558–565.

12. KRUEGER, K.A., T.S. EVANS, S. ECKERT, et al. 2000. Characteristics of hot flushes withraloxifene in postmenopausal osteoporotic women in the MORE trial: Abstr. P-69.North American Menopause Society 11th Annual Meeting. Orlando, FL.

13. NICKELSEN, T., E.G. LUFKIN, B.L. RIGGS, et al. 1999. Raloxifene hydrochloride, a selec-tive estrogen receptor modulator: safety assessment of effects on cognitive functionand mood in postmenopausal women. Psychoneuroendocrinology 24: 115–128.

14. KRUEGER, K.A., K. YAFFE, S. SARKAR, et al. 2000. Effects of raloxifene on cognitivefunction in postmenopausal women without dementia. J. Am. Geriatr. Soc. 48: S106.

15. PEPI WRITING GROUP. 1996. Effects of hormone therapy on bone mineral density:results from the postmenopausal estrogen/progestin interventions (PEPI) trial. JAMA276: 1389–1396.


16. DELMAS, P.D., N.H. BJARNASON, B.H. MITLAK, et al. 1997. Effects of raloxifene onbone mineral density, serum cholesterol concentrations, and uterine endometrium inpostmenopausal women. N. Engl. J. Med. 337: 1641–1647.

17. HULLEY, S., D. GRADY, T. BUSH, et al. 1998. Randomized trial of estrogen plus proges-tin for secondary prevention of coronary heart disease in postmenopausal women.Heart and Estrogen/Progestin Replacement Study (HERS) Research Group. JAMA280: 605–613.

18. GREENDALE, G.A., B.A. REBOUSSIN, A. SIE, et al. 1999. Effects of estrogen-progestinon mammographic parenchymal density. Postmenopausal Estrogen/Progestin Inter-ventions (PEPI) Investigators. Ann. Intern. Med. 130: 262–269.

19. BOYD, N.F., J.W. BYING, R.A. JONG, et al. 1995. Quantitative classification of mammo-graphic densities and breast cancer risk: results from the Canadian National BreastScreening Study. J. Natl. Cancer Inst. 87: 670–675.

20. FREEDMAN, M., J.S. SAN MARTIN, J. O’GORMAN, et al. 2001. Digitized mammography:a clinical trial of postmenopausal women randomly assigned to receive raloxifene,estrogen, or placebo. J. Natl. Cancer Inst. 93: 51–56.

21. PEPI WRITING GROUP. 1996. Effects of hormone replacement therapy on endometrialhistology in postmenopausal women. The Postmenopausal Estrogen/Progestin Inter-ventions (PEPI) Trial. JAMA 275: 370–375.

22. DE VALK-DE ROO, G.W., C.D. STEHOUWER, P. MEIJER, et al. 1999. Both raloxifene andestrogen reduce major cardiovascular risk factors in healthy postmenopausal women.Arterioscler. Thromb. Vasc. Biol. 19: 2993–3000.

23. HEMMINKI, R. & K. MCPHERSON. 2000. Value of drug-licensing documents in studyingthe effect of postmenopausal hormone therapy on cardiovascular disease. Lancet355: 566–569.

24. HERRINGTON, D., D.M. REBOUSSIN, K.B. BROSNIHAN, et al. 2000. Effects of estrogenreplacement on the progression of coronary artery atherosclerosis. N. Engl. J. Med.343: 522–529.

25. LENFANT, C. 2000. Statement on preliminary trends in the Women’s Health Initiative.NHLBI Communications Office; http://www.nhlbi.nih.gov/whi/index.html

26. VISCOLI, C.M., L.M. BRASS, W.N. KERNAN, et al. 2001. A clinical trial of estrogen-replacement therapy after ischemic stroke. N. Engl. J. Med. 345: 1243–1249.

27. MULNARD, R.A., C.W. COTMAN, C. KAWAS, et al. 2000. Estrogen replacement therapyfor treatment of mild to moderate Alzheimer’s Disease: Alzheimer’s Disease Cooper-ative Study. JAMA 283: 1007–1015.

28. JICK, H., L.E. DERBY, M.W. MYERS, et al. 1996. Risk of hospital admission for idio-pathic venous thromboembolism among users of postmenopausal oestrogens. Lancet348: 981–983.

29. PEREZ GUTTHANN, S., L.A. GARCIA RODRIGUEZ, J. CASTELLSAGUE, et al. 1997. Hor-mone replacement therapy and risk of venous thromboembolism: population basedcase-control study. BMJ 314: 796–800.

30. PARSONS, A.K., L.E. NACHTIGALL, P. SULAK, et al. 1999. Vaginal Premarin vs. Replensin women with pre-existing vaginal atrophy receiving oral placebo or raloxifene:effects on subjective endpoints. Menopause 6: 340, P33.

31. PARSONS, A. L. NACHTIGALL, D. MERRITT, et al. 1999. Vaginal Premarin vs. Replens inwomen with pre-existing vaginal atrophy receiving oral placebo or raloxifene:effects on objective endpoints. Menopause 6: 341, P34.

32. COLLABORATIVE GROUP ON HORMONAL FACTORS IN BREAST CANCER. 1997. Breast can-cer and hormone replacement therapy: collaborative reanalysis of data from 51 epi-demiological studies of 52,705 women with breast cancer and 108,411 womenwithout breast cancer. Lancet 350: 1047–1059.

raloxifene: risks and benefits

Documents